test_env: dev
profile: dev-runtime-provider-pool
last_updated_at: "2026-06-28"

source:
  remote_runner:
    ssh: toki@toki-labs.com
    repo_root: /Users/toki/agent-work/iop-dev
    clean_sync:
      - git fetch origin main
      - git reset --hard origin/main
      - git clean -fd
  dirty_policy: discard

edge:
  id: edge-toki-labs-dev
  config_path: build/dev-runtime/edge.yaml
  control_plane_http: http://127.0.0.1:18001
  control_plane_status_url: http://127.0.0.1:18001/edges/edge-toki-labs-dev/status
  bootstrap_http_public: http://toki-labs.com:18082
  openai_base_url_public: http://toki-labs.com:18083/v1
  openai_base_url_runner: http://127.0.0.1:18083/v1
  edge_node_tcp_public: toki-labs.com:18084
  admin_addr_runner: 127.0.0.1:19093

build:
  binaries:
    edge: build/dev-runtime/bin/edge
    node_macos: build/dev-runtime/bin/iop-node
    node_linux_arm64: build/dev-runtime/bin/iop-node-linux-arm64
    node_windows_amd64: build/dev-runtime/bin/iop-node-windows-amd64.exe

model:
  alias: qwen3.6:35b
  provider_capacity_total: 10
  capacity_smoke:
    endpoints:
      - /v1/responses
      - /v1/chat/completions
    concurrent_requests: capacity_plus_one
    expected_total_in_flight: 10
    expected_min_queued: 1
    extended_concurrent_requests: 13
    extended_expected_min_queued: 3
    prompt_policy: long_reasoning_allowed
    exact_output_match: false

nodes:
  - id: mac-codex-node
    alias: mac-codex
    role: cli+mlx-provider
    ssh: toki@toki-labs.com
    workspace: /Users/toki/agent-work/iop-workspace/nomadcode
    provider_pool_candidate: true
    adapters:
      - cli
      - mac-mlx-vllm
    providers:
      - id: mac-mlx-vllm
        type: vllm-mlx
        endpoint: http://127.0.0.1:8002/v1
        served_model: mlx-community/Qwen3.6-35B-A3B-4bit
        capacity: 3
        runtime:
          api_key_policy: local_bearer_from_edge_yaml
          workdir: /Users/toki/agent-work/iop-mlx-vllm
          python: .venv/bin/python
          package_baseline:
            vllm_mlx: 0.3.0
            mlx: 0.31.2
            mlx_lm: 0.31.3
          model_cache: /Users/toki/agent-work/iop-mlx-vllm/hf-cache
          pid_file: /Users/toki/agent-work/iop-mlx-vllm/vllm-mlx.pid
          stdout_log: /Users/toki/agent-work/iop-mlx-vllm/logs/vllm-mlx.stdout.log
          stderr_log: /Users/toki/agent-work/iop-mlx-vllm/logs/vllm-mlx.stderr.log
          max_num_seqs: 3
          max_kv_size: 262144
          max_request_tokens: 262144
          paged_cache_block_size: 64
          max_cache_blocks: 8192
          total_kv_tokens: 524288
          default_chat_template_kwargs:
            enable_thinking: false
          observed_direct_throughput_2026_06_24:
            max_tokens: 192
            completion_token_basis: true
            concurrency:
              1:
                total_tok_s: 59.66
              2:
                total_tok_s: 87.31
              3:
                total_tok_s: 98.13
  - id: gx10-vllm-node
    alias: gx10-vllm
    role: vllm-provider
    ssh: toki@192.168.0.91
    workspace: /home/toki/iop-gx10-vllm
    provider_pool_candidate: true
    provider:
      id: gx10-vllm
      type: vllm
      endpoint: http://192.168.0.91:8001/v1
      served_model: nvidia/Qwen3.6-35B-A3B-NVFP4
      capacity: 4
  - id: onexplayer-lemonade-node
    alias: onexplayer-lemonade
    role: lemonade-provider
    ssh: r0bin@192.168.0.59
    ssh_origin: current_host
    workspace: C:/Users/r0bin/iop-field
    provider_pool_candidate: true
    provider:
      id: onexplayer-lemonade
      type: lemonade
      endpoint: http://192.168.0.59:13305/v1
      served_model: Qwen3.6-35B-A3B-MTP-GGUF
      capacity: 3
      load:
        endpoint: http://192.168.0.59:13305/v1/load
        backend: vulkan
        ctx_size: 786432
        llamacpp_args: "--spec-type none -np 3 -cb -fa on -b 4096 -ub 1024"
        save_options: true
        mtp_runtime_policy: disabled
    windows_process_start: Win32_Process.Create